Expressing Adaptation Strategies Using Adaptation Patterns

HAL Id: hal-00627080

https://hal-supelec.archives-ouvertes.fr/hal-00627080

Submitted on 4 Oct 2011

HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

Patterns

Nadjet Zemirline, Yolaine Bourda

To cite this version:

Nadjet Zemirline, Yolaine Bourda. Expressing Adaptation Strategies Using Adaptation Patterns.

2011. �hal-00627080�

L R I

CNRS – Université de Paris Sud

Centre d’Orsay

LABORATOIRE DE RECHERCHE EN INFORMATIQUE

Bâtiment 490

91405 ORSAY Cedex (France)

R

A

P

P

O

R

T

D

E

R

E

C

H

E

R

C

H

E

EXPRESSING ADAPTATION STRATEGIES

USING ADAPTATION PATTERNS

ZEMIRLINE N / BOURDA Y

Unité Mixte de Recherche 8623

CNRS-Université Paris Sud –LRI

01/2011

Rapport
de
Recherche
N°
1540


Expressing Adaptation Strategies using

Adaptation Patterns

Nadjet Zemirline, Yolaine Bourda, Member, IEEE, Chantal Reynaud

F

Abstract—Today, there is a real challenge to enable personalized access to information. Several systems have been proposed to address this challenge including Adaptive Hypermedia Systems (AHSs). How- ever, the specification of adaptation strategies remains a difficult task for creators of such systems. In this paper, we consider the problem of the definition of adaptation strategies at a high level. We present two main contributions: a typology of elementary adaptation patterns for the adaptation of navigation; and a process to generate adaptation strategies based on the use and the semi-automatic combination of patterns. We also describe how the generated adaptation strategies can be integrated into existing AHSs. A prototype has been implemented and an experiment in the e-learning domain has been conducted with a group of volunteers. This experiment shows that our pattern based approach for defining adaptation strategies is more suitable than those based on ”traditional” AH languages.

Index Terms—AHSs, Adaptation strategies, Patterns.

1 I

NTRODUCTION

The concept of Adaptive Hypermedia Systems (AHSs) has existed for years now [21], and it has amply proved its utility particularly in education [6], [7], where stu- dents have access to personalized resources according to their knowledge, preferences and goals. However, till today, AHSs are not authored as many as desired, and this is mainly due to the difficulty of their authoring process [26].

In fact, authors have to define a domain model struc- turing available resources, a user model describing user characteristics and an adaptation model in the format understood by the used adaptation engine [14]. In this paper, we focus particularly on the authoring process of the adaptation model, which is most often the less intuitive part to be authored in an AHS by non technical persons.

Indeed, authors have to specify an adaptation model, in which they describe resources to propose for users having distinct characteristics and different knowledge in a personalized manner, in order to achieve their specific goals. This is done through the definition of multiple adaptation strategies. By an adaptation strategy, we mean that an author specifies which resources have to

Nadjet Zemirline and Yolaine Bourda are with SUPELEC Systems Sciences (E3S) - Computer Science Department, France

Nadjet Zemirline and Chantal Reynaud are with the Universit Paris-Sud XI, CNRS (LRI) & INRIA - Saclay le-de-France / Projet Leo, France

be proposed and how they will be proposed to a set of users who share the same characteristics. Thereby, authors of an AHS face numerous challenges when defining their adaptation strategies.

The first challenge concerns the expression of adap- tation strategies. Multiple solutions have been pro- posed [13], [15] to make it easier, but they were related to a particular AHS and failed to answer the second and the third challenge, e.g, the author graph tool for AHA! [15]

uses visualization in order to support creators and works only for AHA!.

The second challenge concerns the reuse of adaptation strategies from one system in another one, and the expression of adaptation strategies independently of any AHS. To do so, a new paradigm has been proposed:

”write once, use many” [28]. This paradigm endorses expressing adaptation at a high level, independently of all AHSs and then translating this adaptation into a particular AHS. However, proposed adaptation lan- guages [9], [26], [25] failed to answer the third challenge.

The third challenge concerns the granularity in writing adaptation strategies. It targets to avoid to write several times the common parts of adaptation strategies. To do so, adaptation languages using constructors have been proposed [9], [25], but till today, an adaptation strategy is considered as a whole block and can not be easily reused.

This paper addresses these three challenges. It concen- trates on the ease of defining adaptation strategies at a fine granularity, and on the facility of reusing existing adaptation strategies. In a first time, we focus only on the expression of adaptation strategy for the adaptive navigation, where users are forced to navigate among the proposed navigation paths. This can be either by imposing them a particular order or by recommending them resources [21].

We perceive an adaptation strategy as a combination of elementary parts. Each part corresponds to an ele- mentary adaptation and is bound to a user characteris- tic. A part can belong to different complex adaptation strategies depending on user characteristics. Our work takes up this idea. The notion of elementary adaptation patterns that we propose, is an abstraction of such elementary parts. Elementary adaptation patterns are independent from any application domain, but limited in

a first time to express adaptive navigation. We propose a typology for the elementary adaptation patterns and a semi-automatic process to combine them (the most difficult part is done automatically).

The paper is organized as follows. It presents in section 2 close work on the expression of adaptation, and demonstrates the intuition of our work in section 3 with an example. Section 4 reviews the main aspects of our proposal. Section 5 presents the description of elementary adaptation patterns and their organization in a typology, and section 6 describes how elementary adaptation patterns can be used to define adaptation strategies. In section 7, we discuss how the generated adaptation strategies can be integrated on the top of ex- isting AHSs. Finally, we conclude the paper in section 8.

2 R

ELATED WORKS

Most often, during the authoring process of adaptation on domain and user models, authors ask themselves two questions [3]: what kind of adaptation they can provide for users? and how to produce the desired adaptation? The two questions are answered in that order. For deciding what kind of adaptation they can provide, authors may refer to existing typologies on adaptation (cf. Section 2.1), while for producing adaptation, they need to consider what are the most appropriate adaptation engine and the languages understood by each of them (cf. Section 2.2).

On the other hand, as there are more and more resources available on the web, recent works enable authors not only to define adaptation on their sets of resources but also on those available on the web. So, we present works about integrating adaptive technologies on open corpus (cf. Section 2.3).

2.1 What kind of adaptation could be provided?

The well-known Brusilovsky taxonomy [4] is undoubt- edly the most used typology of adaptation. It describes several methods of adaptation that can be combined to- gether. These methods are organized into three non dis- joint groups: adaptive presentation, content adaptation and adaptive navigation support. This typology relies on the fact that the available resources can be modified and restructured during the adaptation process. Hence, it is not suitable when there is no control of the distributed resources.

As we focus in this paper on the expression of adap- tive navigation, we get a particular interest of methods included in the adaptive navigation support group. The group includes 4 methods:

• direct guidance: supervises users step by step. It is done by proposing to users one link at a time.

• adaptive ordering: defines the priority of all the links of a particular page.

• link hiding and removal: hides, removes or disables links to users (e.g, AHA! [15] hides links that are not relevant to users).

• adaptive link annotation: suggests links to users.

The suggestions are often expressed using visual cues (e.g, WHURLE [24] makes suggestions using colours).

• link generation: creates new links on a page.

2.2 How authors can express their adaptation?

We have grouped existing solutions to express adapta- tion in three main categories.

Adaptation languages accompanied by their adap- tation engine. Adaptation strategies written by these adaptation languages are often expressed in condition- action or event-condition-action rules [15], [24], [22].

However, authoring adaptation using rules is not easy to perform and is time consuming. Thereby, aids have been proposed to make the expression of adaptation easier. E.g, the author graph tool for AHA! [15] which uses visualization in order to support authors: for each new created concept, the tool associates a set of attributes and adaptation rules. Regardless, authors are captive to a particular system. Indeed, adaptation strategies ex- pressed in a system cannot be used outside this system, it has to be rewritten.

Generic adaptation languages accompanied by trans- lators to existing adaptation engines. Some generic lan- guages (independent of any system) have been proposed to specify adaptation [9], [25]. Among them, the LAG language [9], which is an implementation of the specifi- cation of the adaptation language defined in the LAOS model [10]. It includes conversion to the WHURLE [24], Blackboard [1] and AHA! adaptation engines. However, LAG is like a programming language, which is not very suitable for non technical authors (an example is given in section 3). Recently a new Generic Adaptation Lan- guage (GAL¹) has been developed to describe adaptive hypermedia [25]. It argues to gather all functionalities of existing adaptation engines and to be an intermediate language between existing authoring environments and adaptation engines. For that, GAL plans to include trans- lators from existing authoring environments to GAL and from GAL to existing adaptation engines. It describes the navigational structure of a web application using ab- stract constructs (e.g. units, attributes). But, the descrip- tion of adaptation remains difficult to specify, as authors have to write a GAL program (use of SPARQL²queries to select resources) in a sequential way and no aid is proposed for them. Furthermore, generated adaptation strategies by these adaptation languages are considered as a whole block and can not be easily reused.

Hypertext and adaptation patterns. Some design pat- terns for expressing personalization in web applications have been proposed [16], based on commonly used design structures. They are suitable for designers of adaptive systems but not for authors of authors of a

1. GAL is proposed in the context of the GRAPPLE project http://www.grapple-project.org/

2. www.w3.org/TR/rdf-sparql-query/

particular adaptive hypermedia course. Such adaptation design patterns have been proposed in the e-learning domain [12], [19]. Garzotto et al. [19] have proposed pat- terns corresponding to learning styles. Cristea et al. [12]

have proposed a taxonomy of various AEHS (Adap- tive Educational Hypermedia Systems) design patterns according to different learning styles. There is no real formalization and no support for an automatic export to a particular adaptation language. One adaptation strategy (as complex as it can be defined by authors) is expressed using only one pattern. Patterns can not be neither combined together or modified, i.e, authors have to find a pattern corresponding to their desired adaptation strategy, otherwise, they can not express it.

2.3 Open corpus adaptive systems

In the AH community, research concerning the integra- tion of open corpus content into adaptive systems has been under scrutiny for several years - mostly in the field of education [3]. Most of the existing systems are built upon an existing AHS (e.g., [9] on top of [15]). Multiple issues are to be faced in order to develop open corpus based adaptive systems ([12], [20]), including automatic hypertext creation, indexing of open corpus resources and content preparation. None of these systems face the problem of the definition of adaptation, by an AHS author, in a simple way.

2.4 Conclusion

We have discussed here solutions helping authors to find what adaptation they can propose and how they can express it. However, till now, there are no works concerning building complex adaptation strategies, inde- pendent of any system by combining simple adaptations.

In this paper, we focus on this specific point. Adaptation strategies must be defined at a fine granularity. Our aim is thus to help authors defining their own adaptations, independently of any adaptation engine, at a higher level and in an easy manner. In the next section, we introduce a use case giving the intuition of our contribution. This scenario is subsequently used in the paper.

3 M

^OTIVATION

,

^{USE CASE}

Assume that Jane who is a lecturer in computer science wants to build an adaptive course from her materials, i.e., Jane is going to author an AEHS. She has first to de- fine a domain model, then to describe the characteristics of her students in a user model, and finally to express the desired adaptation.

Jane proposes a domain model in UML (cf. Figure 1), in-which she considers the addressed notions as in- stances of the class Concept³. The concepts must be learnt in a particular order, that is defined through the relation

3. In this paper, names of classes have the first letter in upper-case and are in italic, and name of instances have the same name as the class for which they belong in lower-case

Definition Example

Concept abstraction

Resource

Format (Text, Image, Video)

* 1

*

prerequisite

*

Fig. 1. Jane’s domain model

pre-requisite. Each concept may be trained using defini- tions or examples. Definition and Example are subclasses of the class Resource, i.e. each of their instances has a con- tent, which can be proposed to students. Furthermore, each resource may be in different formats: text, image or video.

Jane considers the following student characteristics

• learning mode: in-depth learning mode means that each subject must be known in-depth before going to a related subject. In-breadth learning mode means that a student has to know a variety of subjects before going in-depth.

• reasoning mode: an inductive reasoning mode means that the student has access to examples before the related definitions are presented to him. In a deduc- tive reasoning mode definitions precede examples.

• presentation form: a verbal presentation form is for students preferring textual resources and an audio presentation form is for those preferring audio re- sources.

Among the adaptation strategies Jane wants to pro- pose, we are going to focus on the adaptation strategy S1. It concerns students whose learning mode is in- depth, with an inductive reasoning mode and preferring audio resources. S1 proposes resources that are examples before those which are definitions. They will be in an audio format if that one is available otherwise in a textual format. They will be related to concepts ordered according to a depth-first navigational path using the relation pre-requisite.

Jane can express S1 using solutions supported by her AH system. However, they are not easy to implement and require good backgrounds. See as an illustration, the implementation of S1 using GLAM in figure 2 (for GLAM syntax see section 7.2.1), and using LAG in figure 3. This implies that Jane has already her domain and user models in the format understood by the used AH system.

Naturally, Jane expressed S1 in three parts: 1) S1 con- cerns students whose learning mode is in-depth, 2) S1 concerns students with an inductive reasoning mode, 3) S1 concerns students preferring audio resources. These parts can be considered independently of one another and may compose the definition of other adaptation strategies, for example S2, an adaptation strategy for students whose learning mode is in-depth, with an in- ductive reasoning mode and preferring textual resources.

S2 differs from S1 only in proposing resources in a

S1 is defined in three steps:

Step1: defining GLAM rules

R1 type (r, Example) ∧ format (r, audio) ∧ abstraction(r, Con- cept1) ∧ abstraction(currentR, Concept2) ∧ pre-requisite (Concept2, Concept1) ∧ pre-requisite (Concept1, goal) → Read(r, degree)

R2 type (r, Example) ∧ format (r, text) ∧ abstraction(r, Concept1)

∧ abstraction(currentR, Concept2) ∧ pre-requisite (Concept2, Concept1) ∧ pre-requisite (Concept1, goal) → Read(r, degree) R3 type (r, Definition) ∧ format (r, audio) ∧ abstraction(r, Concept1) ∧ abstraction(currentR, Concept2) ∧ pre-requisite (Concept2, Concept1) ∧ pre-requisite (Concept1, goal) → Read(r, degree)

R4 type (r, Definition) ∧ format (r, text) ∧ abstraction(r, Con- cept1) ∧ abstraction(currentR, Concept2) ∧ pre-requisite (Concept2, Concept1) ∧ pre-requisite (Concept1, goal) → Read(r, degree)

R5 type (r, Example) ∧ format (r, audio) ∧ abstraction(r, Con- cept1) ∧ pre-requisite (Concept1, goal) → Read(r, degree) R6 type (r, Example) ∧ format (r, text) ∧ abstraction(r, Concept1)

∧ pre-requisite (Concept1, goal) → Read(r, degree)

R7 type (r, Definition) ∧ format (r, audio) ∧ abstraction(r, Con- cept1) ∧ pre-requisite (Concept1, goal) → Read(r, degree) R8 type (r, Definition) ∧ format (r, text) ∧ abstraction(r, Con-

cept1) ∧ pre-requisite (Concept1, goal) → Read(r, degree) R1 proposes audio examples according to a depth-first naviga- tional path on their concepts using the relation pre-requisite), and these concepts enable to reach the goal^a.

R5 proposes audio examples that are linked to concepts which enable to reach the goal.

Step2: defining associations between rules and user charac- teristics: γ(audio) is associated to R1, R3, R5, R7, γ(text) is associated to R2, R4, R6, R8, γ(Definition) is associated to R3, R4, R7, R8, γ(Example) is associated to R1, R2, R5, R6, and γ(depth- first) is associated to R1, R2, R3, R4, R5, R6, R7, R8.

Step3: defining GLAM meta-rules MR1 γ(audio) is preferred on γ(text) MR2 γ(Example) before γ(Definition) MR3 R1 before R5

MR4 R2 before R6 MR5 R3 before R7 MR6 R4 before R8

(MR1 means that the rules associated to the characteristic pre- sentation form audio are executed, when they return no results, the rules associated to the characteristic presentation form text are executed).

a. In this paper, the goal to be reached by users (learners for this example) is modelled as a property

Fig. 2. Jane’s S1 in the GLAM format

initialization ( while true (

PM.GM.Concept.show = false UM.Concept.defAudio = false ) while ( enough(GM.Concept.type == Example

GM.Concept.label == audio, 2)) do (PM.GM.Concept.show = true ) )

implementation (

if enough(PM.GM.Concept.access == true GM.Concept.type == Definition , 2)

then (PM.GM.Concept.show = true UM.Concept.defAudio = true)

if enough (PM.GM.Concept.Parent.access == true UM.Concept.defAudio == true

GM.Concept.type == Example , 3) then (PM.DM.Concept.show = true) )

Fig. 3. Jane’s S1 in the LAG format

textual format if that one is available otherwise in an audio format.

To enable Jane easily define her strategies, i.e. the most natural way as possible, we offer the possibility to specify each part of a strategy by defining the set of resources to propose and the order in which they have to be proposed. According to this approach, S1 will be built from the following parts:

S1-1 presents resources linked to the domain concepts ordered according to a depth-first navigational path using the pre-requisite relation.

S1-2 presents only audio resources if they are available otherwise presents textual resources.

S1-3 presents examples before definitions.

The adaptation strategy S1 is intended to students with specific characteristics. Therefore, each part of the strategy has to be labelled by a student characteristic, i.e.

S1-1, for example, will be defined for in-depth learning mode students. Thereby, to define S2, Jane can reuse the parts S1-1 and S1-3, she has only to define the part S2-2 for the textual presentation form.

We presented here the intuition of our contribution according to Jane’s needs, in the following, we describe our approach in a more general way.

4 M

AIN ASPECTS OF OUR FRAMEWORK

We propose the EAP framework in which authors have a clear separation between what kind of adaptation strate- gies they want to provide to users and the technicalities involved in writing it. The idea is to help authors in selecting the adaptation strategy and then generated it in a semi-automatic way. Defined adaptation strategies are described at a high level and independently of any adaptation engine.

The EAP framework focuses only on the expression of adaptation strategies. So, it assumes authors have already created their domain and user models. Further- more, our framework is based on design patterns [18].

Design patterns describe recurrent solutions to common problems in software design. The solutions are generic and cannot be directly translated to code. For e.g, the Object-oriented design patterns describe relationships and interactions between classes or objects, without specifying the final application classes or objects that are involved. In practice, design patterns can speed up the development process by providing tested, proven development paradigms. We argue that an adaptation strategy is a kind of conception, where authors have to write several times the same parts of an adaptation strategy, sometimes on different elements. Consequently, the proposed framework uses a set of building blocks independent from any application domain, called ele- mentary adaptation patterns, which are based on design patterns. Thereby they can be used and instantiated to define specific adaptation strategies.

The main steps for authoring an adaptation strategy with the EAP framework are:

1) Selection. The author either selects elementary adaptation patterns (those needed to define his adaptation strategy) and instantiates them on his own model (thereby, elementary adaptations are defined), or reuses existing elementary adaptations.

2) Specification. The creator specifies associations be- tween user characteristics and elementary adapta- tions.

3) Computation. The computation of the adaptation strategy resulting from step 2 is automatic.

We have defined a typology and a library of elemen- tary adaptation patterns that can be selected for use within an adaptation strategy, and which we introduce in section 5. The instantiation process and the combination process are described in section 6.

Before going further, let us apply the EAP framework on Jane’s use case. Jane needs now:

1) To define S1-1 (resp. S1-2, S1-3) by instantiating the appropriate elementary adaptation pattern on the relation pre-requisite (resp. on the classes Example, Definition, on the property format).

2) To associate S1-1 with in-depth learning mode, S1- 2 with inductive reasoning mode, S1-3 with audio presentation form.

3) As S1 is for students with an in-depth learning mode, an inductive reasoning mode and who want audio resources. S1 is thus automatically built by combining S1-1, S1-2, S1-3.

Note that, that way, Jane does not need to worry about technical problems in the expression of S1.

5 E

LEMENTARY ADAPTATION PATTERNS The notion of elementary adaptation patterns that we propose, is an abstraction of Jane parts. Furthermore, we defined our elementary adaptation patterns in a manner that is independent from any application domain in order to be able to cover other authors parts. Thereby, the criteria used to define our elementary adaptation patterns are defined in a generic way (cf. Section 5.1).

Elementary adaptation patterns are described in sec- tion 5.2, and their typology is defined in section 5.3.

5.1 Fundamental criteria for defining elementary adaptation patterns

As each part of the S1 strategy defined by Jane, an elementary adaptation pattern targets a set of resources of a particular type to be presented and also specifies the order in which they will be proposed. This section presents exhaustive criteria to select resources (cf. Sec- tion 5.1.1) and to organize the selected resources (cf.

Section 5.1.2).

5.1.1 Criteria used to select resources

Criteria used to select resources are based on the domain model, where resources are structured and described. We

argue that the general description of a domain model includes the following elements:

• a set of classes. This set must contain the class representing all the resources to be proposed to users which we have called Resource, and the class representing all the domain concepts, which we have called Concept.

• a set of relations between classes. Each relation defines a graph on instances of classes on which it is de- fined. This graph can be navigated according to two different navigational paths in order to reach the goals: depth-first or breadth first.

• a set of properties.

Thereby, we have differentiated between criteria select- ing resources and criteria defining a navigational path on relations. Our criteria for selecting resources are: their belonging to a class, the values of some properties, or the presence of a relation that defines a navigational path through the resources or the concepts graph. Fur- thermore, our criteria currently considered for defining a navigational path are either depth-first, breadth-first or random.

5.1.2 Criteria used to order the selected resources We have looked over works defining adaptation meth- ods, by giving a particular interest for adaptive naviga- tion, without mattering if the methods are applied on a set of links to resources or resources themselves.

We have looked over the Brusilovsky typology (cf. Sec- tion 2.1) excluding methods of the adaptive navigation support group which modify resources (e.g, hiding links belonging to content of resources). Only direct guidance, adaptive ordering and adaptive link annotation have been considered.

We have also looked over the classification of external actions in AHS defined by Stash and al. [26]. The classi- fication includes actions on items (e.g, selection, showing items or links to items), actions on a set of items (e.g, ordering), hierarchical actions (e.g, action on parent or child) and actions on the overall environment (e.g, changing the layout). We only consider the actions having impact on the navigation of users, this includes: actions on items, actions on a set of items and hierarchical actions. Further- more, we distinguish between actions and elements on which the actions are performed. The elements can be an item, a set of item, parents or children. So, we only consider the selection, show and order actions.

On the other hand, we have looked over AHS implementing adaptive navigation like AHA! [15], WHURLE [24], GLAM [22] etc. We found that GLAM implements a kind of adaptation not mentioned else- where. This adaptation proposes alternative resources if the desired resources are not available. We find it interesting and have retained it in our own typology.

From this study, we conclude that there are four basic modes to select resources in a setting of adaptive navigation support. These modes are the following:

• Selection only: provides a set of resources, which are all proposed to the user, i.e, only the selected resources are proposed to users, the other resources are not proposed. This selection mode is equivalent to the combination of the selection and show actions described by Stash et al. In fact, Stash et al. propose to select and show selected resources in two sepa- rate processes, while in our approach implicitly all selected resources are shown. There is no equivalent in the Brusilovsky typology.

• Recommended selection: provides multiple sets of re- sources (at least two) that include knowledge to specify which set should be recommended rather than the other (sets of) resources. for example, we can recommend definitions rather than examples.

The user can access both types of resource, but a typographic indication enables the user to iden- tify which resources are recommended. It is equiv- alent to the adaptive link annotation described by Brusilovsky, but there is no equivalent in the actions described by Stash et al.

• Ordered selection: provides multiple sets of resources (at least two), accompanied with knowledge to spec- ify the order in which they must be presented. Only one set of resources is proposed at a time, and the resources of a particular set are not proposed until all the resources of all sets of higher priority have been viewed by the user. E.g, in e-learning, concepts can be selected and ordered using the pre-requisite relation defined between concepts. It is equivalent to the adaptive ordering described by Brusilovsky and to the direct guidance described by Brusilovsky when the returned result includes only one resource in each set. It is also equivalent to the combination of the order and show actions described by Stash et al..

• Alternate selection: provides multiple sets of re- sources (at least two), accompanied with data that specifies the order in which they must be presented, knowing that only one set is presented to the user.

E.g, we propose textual resources when they are available, and audio resources in the absence of textual resources. Neither Brusilosvky or Stash et al.

has considered this selection mode.

5.2 Description of elementary adaptation patterns We propose the following definition for elementary adaptation patterns, based on the definition of design patterns [18].

Definition 1: An elementary adaptation pattern describes a generic solution for a generic elementary adaptation problem.

This solution is independent from any language, and it exploits the characteristics of the domain model.

Definition 2: A generic elementary adaptation problem de- scribes a criterion to select resources to be proposed and a criterion to define in which order the selected resources are going to be proposed.

Name: the name of the elementary adaptation pattern described.

Intent: the intent is a short statement about an elementary adaptation problem. It answers the following questions: what is the elementary adaptation pattern supposed to do? i.e. what is its goal? Indeed, it indicates the way the resources are selected and the way they are presented.

Solution: the solution includes two elements:

Expressions: denote a set of resources to be proposed to the user, and the conditions which have to be satisfied. These conditions can be represented in one or more logical expres- sions. Those to be considered simultaneously are gathered in the same expression, while excluded conditions are expressed in different expressions. The formal description of expressions may be accompanied by an informal description.

Meta-expressions: a binary relation between two expressions.

Indeed, when using multiple expressions, we specify the way they have to be considered by using meta-expressions. The formal description of meta-expressions may be accompanied by an informal description.

Constituents: describe the elements of the domain model used in the expressions described in the solution pattern.

Fig. 4. Description of elementary adaptation patterns

We define in the figure 4 the characteristics retained from [18] and used to describe elementary adaptation patterns.

The solution part is the most formal part of the ele- mentary adaptation patterns. We have defined a gram- mar using the Extended Backus-Naur Form (EBNF) [29].

The grammar is described in the figure 5. It includes a set of non-terminal elements expressed between brackets, and a set of terminal elements expressed between coats.

For people not familiar with EBNF syntax, we give examples of the solution part respecting the proposed grammar (cf. Figure 8, 9, 10). These examples are also accompanied by an informal description.

We give an informal description of the semantic of the language defined by the grammar and some associated constraints. In order to do so, we consider a domain model DM, composed of:

• Cls = {c/ c is a class}

• Rel = {rel/ rel is a relation}

• Prop = {p/ p is a property}

• Valp = {v/ v is a value of the property p}

• Res = {r/ r is a resource}

We defined general elements, which we describe in figure 6. Furthermore, we defined predicates to facilitate the selection of either resources or concepts. These pred- icate are:

• instanceOf : instanceOf(r, c) is true, for all resources r that are instances of the class c.

• characteristicOf : characteristicOf(r, p, op, v) is true, for all resources r having the property p and satisfying the comparison test using the operator op and the value v.

• linked: linked(i1, i2, rel) is true, for all instances i1 that are linked directly to the instance i2 by the relation rel.

• linked-transitive: linked-transitive(i1, i2, rel) is true, for all instances i1 that are linked directly or indirectly to the instance i2 by the relation rel.

hSolutioni ::= hExpressionsi hMeta-Expressionsi.

hExpressionsi ::= (hExpressionreli)*| (hExpressionpropi)*|

(hExpressionclsi)*.

hMeta-expressionsi ::= (hIdi ”≺” hIdi)* | (hIdi ”]” hIdi)* | (hIdi

”|” hIdi)*.

hExpressionreli ::= hIdi ”:”hExpreli ( ”∧” hExpreli )*.

hExpressionpropi ::= hIdi ”:”hExppropi.

hExpressionclsi ::= hIdi ”:”hExpclsi.

hExpreli ::= linked ”(”hInsti”,” hInsti ”,” hReli ”)” | linked-transitive ”(”hInsti”,” hInsti ”,” hReli ”)” | distance ”(”hInsti”,” hInsti ”,” hReli ”,” hNumberi”)”.

hExppropi ::= characteristicOf ”(”hResi”,” hPropi ”,” hOperatori

”,” hVali”)” .

hOperatori ::= ”=” |”6=” | ”≤” | ”≥”.

hExpclsi ::= instanceOf ”(” hResi”,” hClsi ”)” . hIdi ::= hStringi.

hClsi ::= ”c”hNumberi .

hInsti ::= ”concept”hNumberi | hResi.

hResi ::= ”resource”hNumberi .

hReli ::= ”r”hNumberi . hPropi ::= ”p”hNumberi . hVali ::= (hStringi — hNumberi)+.

hStringi ::= [ ”a”-”z” ] hStringi * . hNumberi ::= [ ”0”-”9” ] hNumberi * .

Fig. 5. Syntax of the characteristic Solution

Elements Variable referring to

<Number> any integer number

<String> any string

<Id> identifiers. Identifiers belonging to the same solution part have to be different

<Res> a resource

<Inst> either a concept or a resource

<Cls> a class of DM

<Rel> a relation of DM

<Prop> a property of DM

<Val> a value among the allowed values for the used property

Fig. 6. Description of general elements

• distance: distance(i1, i2, rel, n) is true, for all instances i1 that are distant from the instance i2 by n instances using the relation rel.

These predicates compose 3 types of expressions:

• <Expcls>for expressions on classes.

• <Expprop> for expressions on properties. Expres- sions belonging to the same solution part have to be expressed on the same property.

• <Exprel> for expressions on relations. When the expression includes multiple selections, the vari- ables indicating the selected resources have to be the same.

When more than one expression is defined in a so- lution, meta-expressions must be defined between all expressions of the solution. This is done using the ex- pression identifiers. Each identifier used in the definition of a meta-expression must correspond to an expression identifier. Three types of meta-expressions are proposed.

They are:

• <Id1> ≺ <Id2> means that the set of resources selected with the expression identified by Id1 is proposed before the set of resources selected with the expression identified by Id2.

• <Id1> ] <Id2> means that the set of resources selected with the expression identified by Id1 is rec- ommended rather than the set of resources selected with the expression identified by Id2. A typographic indication can be used to differentiate between the set of resources recommended from those they are not.

• <Id1> | <Id2> means that the set of resources selected with the expression identified by Id2 is an alternative of the set of resources selected with the expression identified by Id1.

5.3 Typology of elementary adaptation patterns We have defined a library of 22 elementary adaptation patterns using the criteria defined in section 5.1. An elementary adaptation pattern is based simultaneously on (1) one of the 4 selection modes of resources to be proposed, (2) one of the 3 elements of the domain model involved in the selection process and when the element is a relation, we consider also (3) one of the 2 types of navigation through the resources or the concepts graph. The two navigation modes are applied for all the selection modes except for the selection only mode, which proposes a set of resources according to a particular criterion.

In order to be able to look easily over the defined elementary adaptation patterns, we have organized them in a tree where each leaf is an elementary adaptation pattern (cf. Figure 7). The tree represents our typology.

Let us now use this typology to help Jane to define S1. We note that each part of S1 can be defined thanks to a pattern. The pattern P2.1.1.1 (cf. Figure 8) is used to define S1-1 (S1-1 consists of ordering concepts according to a depth-first navigational path using the relation pre- requisite, and presents resources linked to these con- cepts), P3.3 (cf. Figure 9) is used to define S1-2 (S1-2 consists of presenting only audio resources if they are available otherwise presents textual resources), and P2.2 (cf. Figure 10) is used to define S3-3 (S3-3 consists of presenting examples before definitions).

After having described the typology and some elemen- tary adaptation patterns, lets come back to the process of defining adaptation strategies.

6 D

EFINING ADAPTATION STRATEGIES

This sections focuses on the steps 1 and 3 of the au- thoring steps of adaptation strategies (cf. Section 4).

Adaptive navigation

1. Selection only

2. Ordered Selection

3. Recommended Selection

.1. Relations .2. Classes

.1. Relations .2. Classes .3. Properties

.2. Classes .1. Relations

.2. Classes .3. Properties

.3. Properties .3. Properties Type of navigation on

the domain model

Selection modes Elements of the domain model

4. Preferred Selection

..2. Resource

.1. Relations

..1. Concept ..1. Concept

..2. Resource

...2. Breath-first ...1. Depth-first

...2. Breath-first ...1. Depth-first Classes related

to relations

Navigational path on instances Patterns

P 1.1.1 P 1.1.2

P 1.3 P 1.2

P 2.1.1.2 P 2.1.1.1

P 2.1.2.2 P 2.1.2.1

P 2.3 P 2.2

P 3.3 P 3.2 ..1. Concept

..2. Resource

...2. Breath-first ...1. Depth-first

...2. Breath-first ...1. Depth-first

P 3.1.1.2 P 3.1.1.1

P 3.1.2.2 P 3.1.2.1

P 4.3 P 4.2 ..1. Concept

..2. Resource

...2. Breath-first ...1. Depth-first

...2. Breath-first ...1. Depth-first

P 4.1.1.2 P 4.1.1.1

P 4.1.2.2 P 4.1.2.1

Fig. 7. Typology of elementary adaptation patterns

Name: Ordered Selection - Depth first- Relation - Concept Intent: This pattern proposes resources according to a depth first navigational path on concepts.

Solution:

Expression

E1: linked(rCurrent, concept’, abstraction) ∧ linked- transitive(concept, goal, relationi) ∧ linked(r, concept, abstraction)

∧ linked(concept, concept’, relationi)

E2: linked-transitive(concept, goal, relationi) ∧ linked(r, concept, abstraction)

According to E1: selected resources are linked to concepts using abstraction. these concepts can reach the goal using relationi and are directly linked to the current concept. Ac- cording to E2: selected resources are linked to concepts using abstraction, these concepts can reach the goal using relationi.

Meta-expressions E1≺ E2

According to this meta-expression, the set of resources selected by E1is proposed before the set of resources selected by E2.

Constituents:

concept: a variable describing an instance of the class Con- cept.

rCurrent: a variable describing the current instance pro- posed to users of the class Resource or of one of its special- izations.

goal: a variable describing the goal to reach, which is an instance of the class Concept.

r: a variable describing an instance of the class Resource or of one of its specializations.

relationi: a variable describing a relation defined between instances of the class Concept.

abstraction: a variable describing a relation defined between an instance of the class Concept and one or more instances of the class Resource or of one of its specializations.

Fig. 8. OrderedSelection-DepthFirst-Relation-Concept

Name: Ordered Selection - Classes

Intent: This pattern proposes ordered resources belonging only to subclasses of the class Resource.

Solution:

Expressions

E1: instanceOf (r, Class1) ...

En: instanceOf (r, Classn)

According to Ei: selected resources are instances of the class Classi

Meta-expressions

Ei≺ E_j, i < j, i = 1..n and j = 1..n.

According to this meta-expression, the set of resources selected by Ei is proposed before the set of resources selected by Ej(i < j).

Constituents:

r: a variable describing an instance of the class Resource or of one of its specializations.

Classi: a variable describing a subclass of the class Resource.

Fig. 9. Ordered Selection-Classes

Name: Alternate Selection - Properties

Intent: This pattern proposes resources that satisfy particular values of the property propertyi, if no resources are avail- able, other resources that satisfy other values of the property propertyiwill be proposed.

Solution:

Expressions

E1: characteristicOf(r, propertyi, op, val1) ....

En: characteristicOf(r, propertyi, op, valn)

According to Ei: selected resources have the property propertyiand their value must satisfy the comparison test.

Meta-expressions

Ei| Ej, i < j, i = 1..n and j = 1..n.

According to this meta-expression, the set of resources selected by Ejis an alternative of the set of resources selected by Ei(i < j).

Constituents:

r: a variable describing an instance of the class Resource or of one of its specializations.

propertyi: a variable describing a property of the class Resource.

val: a variable describing a possible value for the property propertyi.

Fig. 10. Alternate Selection-Properties

The step2 is simple and we do not give further details.

We start first by describing the step1 related to the instantiation process of elementary adaptation patterns (cf. Section 6.1). Then, we detail the step3 related to the combination process (cf. Section 6.2), and we end by using the EAP framework to define Jane’s adaptation strategy S1 (cf. Section 6.3).

6.1 Defining elementary adaptations

In order to propose a generic solution, elementary adap- tation patterns are defined on a generic domain model.

Consequently, when authors select an elementary adap- tation pattern, they have to instantiate its constituents on their personal domain model, in order to obtain the elementary adaptation that meets their needs. We define elementary adaptations as follows:

Definition 3: An elementary adaptation is obtained after an instantiation of an elementary adaptation pattern on a

Name: Ordered Selection-Example-Definition

Intent: This elementary adaptation proposes ordered resources belonging only to Example and Definition in this order.

Solution:

Expressions

E1: instanceOf (r, Example) E2: instanceOf (r, Definition)

According to E1, selected resources are instances of the class Example, and according to E2, selected resources are instances of the class Definition.

Meta-expressions E1≺ E2

According to this meta-expression, all examples are pro- posed before all the definitions.

Constituents:

r: a variable which represents an instance of the class Resource or of one of its specializations.

Example: a variable which represents the class Example, a subclass of the class Resource.

Definition: a variable which represents the class Definition, a subclass of the class Resource.

Fig. 11. The elementary adaptation S1-3

particular domain model.

Elementary adaptations has therefore the same struc- ture as elementary adaptation patterns. The generation of an elementary adaptation is done in a semi-automatic way: the characteristics Name, Intent are generated in a semi-automatic way and the characteristics Solution and Constituents are automatically generated.

For example, when Jane wants to express the S1-3 part, she selects the pattern P2.2 and instantiates it with the classes Example and Definition (for informal description, see section 3, for formal description see figure 11).

Following this principle, authors can define several elementary adaptations, each of them being associated to one user characteristic (step2 in section 4). When users have a profile composed of several characteristics, complex adaptation strategies have to be defined. They are obtained by combining elementary adaptations, each one being associated with a component of the user profile. This combination process is detailed in the next section.

6.2 Combining elementary adaptations

Combining elementary adaptations together defines a combined adaptation.

Definition 4: A combined adaptation defines a set of re- sources that satisfy simultaneously all constraints imposed by multiple elementary adaptations.

A combined adaptation has the same characteristics and is structurally identical to an elementary adaptation.

Concretely, the combination process of a set of elemen- tary adaptations consists in combining their character- istics together. A manual process is used to combine the characteristics Name and Intent as it needs natural language processing (not detailed in this paper). We propose an automatic process to combine the character- istics Solution and Constituents which is explained further below.

The combination of the characteristic Constituents is simple. Constituents coming from the different adapta- tions are gathered together into a set of constituents.

However, the combination of the characteristic Solution is more complex and we have defined the following process.

We have chosen to base the process on criteria con- cerning the selection of resources, as our final aim is to propose a set of resources. Thereby, we have criteria based on: classes to which a resource belongs, proper- ties satisfied by a resource, and relations in which a resource participate. We have exploited these criteria in the combination process of the characteristic Solution.

We express this process in two sequential steps:

1) Build different sets of identifiers of expressions, one set for each different criterion (cf. Section 6.2.1).

2) Build one adaptation from the sets built in step 1 (cf. Section 6.2.2).

6.2.1 Step one of the combination

Let Sol1, Sol2, ..., Soln be the solution part of the el- ementary adaptations to combine, where each Soli is composed of:

• ni expressions noted Ei, each expression having an identifier Idi.

• mi meta-expressions noted MEi.

We group the identifiers whose expressions are ex- pressed on one given criteria in different sets.

• the identifiers whose expressions exploit classes are put in the same set Setcls= {Idi/ Idi is an identifier that denotes an expression exploiting classes}.

• the identifiers whose expressions exploit relations are grouped into sets, one set per relation. Setrel = {Idj/ Idj is an identifier that denotes an expression exploiting the relation rel}.

• the identifiers whose expressions exploit properties are grouped into sets, one set per property. Setprop= {Idj/ Idj is an identifier that denotes an expression exploiting the property prop}.

where each Idi ∈ Soli belongs only to one set, either to the Setcls, to a set of {Setrel}, or to a set of {Setprop}.

6.2.2 Step two of the combination

Let Set1, Set2, ..., Setp be the sets of identifiers obtained after the first step, let Solcbe the solution resulting from the second step of the combination process composed of:

• n_j expressions noted CEc.

• m_j meta-expressions noted CMEc.

Let Setc be the set of p tuples built as follows:

Setc = Set1XSet2X...XSetp

For each tuple, a distinct identifier is defined and is associated to an expression CEc:

CE_c= E₁∧ E₂... ∧ E_p where

• CE_c is the expression belonging to Solc.

• E_i is the expression whose identifier is Idi, and Idi∈ Soli, i = 1...p.

Identifiers are also used to associate knowledge to expressions. This results in defining meta-expressions.

Defining meta-expressions on the expressions Ec of the solution Solc is done as follows.

Let CEiand CEj be two expressions belonging to the solution Solc, where CEi (resp. CEj) contains E1 (resp.

E2), E1 and E2 belong to the same solution, and are linked by the meta-expression Id1MhId2(Id1(resp. Id2) is the identifier of E1(resp. E2) ). In that case, we deduce the meta-expression Idi Mh Idj (where Idi (resp. Idj) is the identifier of CEi (resp. CEj) ).

However, as a meta-expression is an anti-symmetric binary relation between two expressions, two types of conflict can be encountered. They are processed auto- matically (by deleting all meta-expressions in conflict except one). The process uses a default solution that can be changed by the author.

• Conflict 1: The generation of the same relation be- tween CEi and CEj and between CEj and CEi

(e.g. CE1 ≺ CE2 and CE2 ≺ CE1). We propose to order sets of adaptations obtained after the first step according to (1) sets based on the navigational path of the graph, (2) sets exploiting the type of the resources, (3) sets exploiting the characteristics of the resources.

• Conflict 2: The generation of two meta-expressions between two identical expressions (e.g. CE1≺ CE2

and CE1 ] CE2). We give a different priority to meta-expressions according to the defined relation:

(1) Priority, (2) Recommendation, (3) Alternate.

We have implemented the following deduction pro- cess of CM Ec. The p sets of identifiers coming from the first step are first ordered according to the proposed or- der in the resolution of conflict 1. In a second time, each meta-expression defined using these identifiers allows us to deduce multiple meta-expressions of CM Ec. Each time a meta-expression is deduced, we check if it does not generate a conflict with the already generated meta- expressions. If a conflict of the first type is generated, the current meta-expression is not considered and the deduction process will continue. If a conflict of the second type is generated, we retain only one meta- expression according to the order defined in the solution of the second conflict.

6.3 Jane’s adaptation strategy

We apply here our framework in order to define Jane’s adaptation strategy S1. We consider that the elementary adaptation S1-1, S1-2 and S1-3 (cf. Figure 12) have been defined. Then, Jane has established correspondences be- tween each elementary adaptation and a user character- istic S1-1 with in-depth learning mode, S1-2 with inductive reasoning mode, S1-3 with audio presentation form. We

Expressions Meta-

expressions S1-1 E1−1 = linked-transitive(r, goal,

prerequisite) ∧ linked(rCurrent, r, prerequisite)

E1−1≺ E1−2

E1−2 = linked-transitive(concept, goal, pre-requisite) ∧ linked(r, con- cept, abstraction)

S1-2 E2−1 = characteristicOf(r, format,

=, audio) E2−2=characteristicOf(r, format, =, text)

E2−1| E2−2

S1-3 E3−1 = instanceOf(r, Example) E3−2=instanceOf(r, Definition)

E3−1≺ E3−2

Fig. 12. Description of S1-1, S1-2, S1-3

focus now on the way S1-1, S1-2 and S1-3 are combined in order to produce S1.

Next, Jane associates S1-1 with in-depth learning mode, S1-2 with audio display mode, S1-3 with inductive reason- ing mode.

S1-1, S1-2 and S1-3 are combined automatically to define S1. The combination process of their characteristic Solution is performed as follows. It has as input 3 ele- mentary adaptations expressed on 3 different elements of the domain model. After the step 1 of the combination process, 3 sets are built, one adaptation per set. After the step 2, one combined adaptation is built, which is composed of 8 expressions and 44 meta-expressions.

Among the deduced expressions, we have:

• Ec,1=E1−1∧ E2−1∧ E3−1=linked-transitive(r, goal, prerequisite) ∧ linked(rCurrent, r, prerequisite) ∧ characteristicOf(r, format, =, audio) ∧ instanceOf(r, Example)

• Ec,2=E1−1∧ E2−1∧ E3−2=linked-transitive(r, goal, prerequisite) ∧ linked(rCurrent, r, prerequisite) ∧ characteristicOf(r, format, =, audio) ∧ instanceOf(r, Definition)

• Ec,3=E1−2∧ E2−1∧ E3−1=linked-transitive(r, goal, prerequisite) ∧ characteristicOf(r, format, =, audio)

∧ instanceOf(r, Example)

• Ec,4=E1−2∧ E2−2∧ E3−1=linked-transitive(r, goal, prerequisite) ∧ characteristicOf(r, format, =, text) ∧ instanceOf(r, Example)

• Ec,5=E1−2∧ E2−2∧ E3−2=linked-transitive(r, goal, prerequisite) ∧ characteristicOf(r, format, =, text) ∧ instanceOf(r, Definition)

Among the deduced meta-expressions retained, we have: Ec,1 ≺ Ec,2, Ec,2 ≺ Ec,3, Ec,2 ≺ Ec,4, Ec,2 ≺ Ec,5

and Ec,3| Ec,4.

Among the deduced meta-expressions not retained, we have: Ec,3 ≺ Ec,2, Ec,4 ≺ Ec,2, Ec,1 | Ec,4, Ec,2 | Ec,5

and Ec,2| Ec,4.

7 V

^ALIDATION

In this paper, we promote two main ideas behind the EAP framework: enabling authors to specify their adap- tation strategies at a high level, and easiness of defining authors’ adaptation strategies. Here, we prove these ideas by presenting the implementation of our framework (cf.

Section 7.1), by discussing the execution of generated adaptation strategies using an existing adaptation engine (cf. Section 7.2) and by evaluating the expression of adaptation using the EAP framework versus a rule-based language (cf. Section 7.3).

7.1 Implementation of the EAP framework

The framework has been implemented as a plug-in of the Prot´eg´e tool⁴, called EAP. Currently, EAP plug-in is under test. Its architecture is presented in section 7.1.1 and its main functionalities are described in section 7.1.2.

7.1.1 Architecture of the EAP plug-in

As described in Figure 13, the plug-in includes two parts.

First, a knowledge part gathers the library of ele- mentary adaptation patterns and combination rules. The library is modelled in OWL⁵, where each elementary adaptation pattern is an OWL class and is defined as a specialization of a class called ElementaryAdaptation- Pattern. On the other hand, the combination rules im- plement the combination process (cf. Section 6.2) in a declarative way using SWRL rules⁶ and the swrlx built- ins⁷.

Second, the process part is made of components per- forming interaction with an inference engine (in our case Jess) and the OWL Prot´eg´e editor. We have used the OWL Prot´eg´e API to manipulate the creator’s domain and user models, the library of elementary adaptation patterns and their instantiations. We have also used the SWRL Jess Bridge to execute SWRL rules using Jess.

OWL editor (Protégé v 3.4.2)

Inference Engine (Jess ) AH

Author

EAP plug-in Knowledge part

Rule base (SWRL) Elementary

Adaptation Patterns (OWL)

Processing Part Protégé

(OWL API)

SWRL Jess Bridge

Fig. 13. Architecture of the EAP plug-in

7.1.2 Interaction with the EAP plug-in

The plug-in proposes multiple facilities. The author starts by loading his user and domain models. He can then define elementary adaptations by selecting an ele- mentary adaptation pattern, and the constituents of the elementary adaptation. The solution part will then be generated automatically. The author can later define as- sociations between an elementary adaptation and a user characteristic, while the combination process of multiple elementary adaptations is done automatically. Finally,

4. http://protege.stanford.edu/

5. www.w3.org/TR/owl-guide/

6. www.w3.org/Submission/SWRL/

7. The swrlx built-ins augment swrl rules with additional function- alities, e.g, creating new instances

EAP helps authors to export their adaptation strategies automatically in the GLAM format. This conversion is done automatically and is described in the following section. We plan to implement additional extensions, for example, to be able to generate LAG adaptation.

7.2 Execution of generated adaptation strategies Adaptation strategies generated using the EAP frame- work are expressed at a high level, and are indepen- dent of any adaptation engine. Therefore, translators to existing adaptation engines are needed to execute these adaptation strategies. In this paper, we present our work to plug our framework on the GLAM platform, in order to be able to execute generated adaptation strategies by the GLAM adaptation engine. Before presenting this process, we first describe the GLAM platform.

7.2.1 GLAM platform

GLAM (Generic Layered Adaptation Model) is a plat- form defined for an entire class of adaptive hypermedia systems. The platform is made up of a generic adaptation model relying on generic user and domain models. Spe- cific systems can be obtained by specializing the GLAM generic user and domain models. An adaptation strategy in GLAM is described in two levels:

A level based only on domain-related knowledge. It concerns data about the domain model and the position of the user in the domain model. It is exploited using rules. Rules are expressed using a condition-conclusion format as:

predicate1∧... ∧ predicaten→ Action (resourcei, degree)

The condition part describes the conditions having to be satisfied by resources proposed to users. Usually, this part is related to the existence of a relation defining a particular navigational path in the domain model, even- tually to a type of resources or to restrictions concerning the resource format expressed using attributes of the Concept or Resource classes.

The conclusion part describes the activity proposed to users for proposed resources. It includes two elements:

• Action: describes the proposed activity for the pro- posed resource (resourcei in the rule above).

• Degree: can be used in different treatments. In GLAM, it is used to describe the relevance of a re- source against the others. It allows several resources to be proposed to the user, the degree of relevance being represented with a code (color for example).

The degree of relevance has five values (very high, high, medium, low, and very low), each value is associated to a particular color.

A level based on user-related knowledge and user characteristics. It is exploited using meta-rules.

Meta-rules describe mechanisms that govern selection, scheduling, and excluding rules for a given user accord- ing to his profile. Let R1, R2 be two sets of rules, where